Automated navigation systems are still wreaking havoc on small towns’ streets
Apps like Google and Waze are redirecting traffic to secondary roads that are not equipped to handle the traffic, disrupting their infrastructure. Small cities have little capacity to change this.
Before the pandemic, there were plenty of news articles on one specific topic in France. In Tremuson, in the West of the country, in Civrieux d’Azergues, near the city of Lyon, in Lieusaint or Aulnay-sous-Bois, in the vicinity of Paris, the same problem occurred: traffic was displaced from main to secondary roads.
This contributed to the degradation of these smaller roads, which were not built to handle heavy traffic. It also created new dangers for pedestrians and residents of the neighborhoods where so many cars were rerouted. For many, it was obvious: all these disruptions were the fault of GPS applications like Waze, Google Maps, or Apple Plans.
In calculating the fastest route from one point to another, these tools often suggest drivers take shortcuts through residential areas or secondary routes. To stop the digitally enhanced congestion, many mayors and traffic managers thought that the best solution would be to tweak the apps so that they would only suggest highways and main roads. But of course, outside the offices of Google and Apple, nobody exactly knows how these algorithms calculate the best routes for their users. (Google bought Waze for $1.3 billion in 2013. They recently announced the merger of Google Maps and Waze’s services teams, which will result in a reduction of Waze monetization-focused roles).
Then, Covid-19 hit, followed by lockdowns. “This definitely had an impact on the number of traffic jams,” says Camille Combe, co-author of a 2019 report on the end of urban congestion. “It also showed that rather than using GPS applications, a solution to fight congestion is simply to reduce the number of cars on the roads.”
In 2023, however, traffic is coming back to its pre-pandemic level. And with it, a set of questions: are GPS apps the real culprit for secondary road congestion? If so, is there a way to fight them? In France, a first-of-its-kind decree was passed in August 2022 to “avoid encouraging massive use of secondary roads for through traffic”. Does it change anything in the navigation apps’ calculations?
Collective versus individualistic approach
To understand the root of the problem, it is useful to study infrastructure managers’ and GPS applications’ divergent approaches to traffic. In a 2018 article, the sociologist Antoine Courmont described how public policies regulate traffic from a global point of view, trying to provide an infrastructure where road flows create demand. GPS applications, however, take a more individualistic approach: they consider traffic jams as a problem that adversely impacts each driver.
As such, navigation applications only try to improve the individual outcome for each driver and do not take into account the traffic plans drawn up by local authorities. In 2018, another study by the UC Berkeley Institute of Transportation Studies showed how, in case of congestion, just 20% of drivers following app suggestions could create new traffic jams further up the main roads as well as on the alternative routes they would take.
Five years on, in Tremuson, a small town of fewer than 2,200 inhabitants close to the bigger Saint-Brieuc (45,000 inhabitants) and to the main and often congested road RN 12, this pattern of accidental traffic jam creation still shows. “As soon as delays appear on the expressway, the applications suggest to drivers to go through the center of the town,” says Mayor Yves Orgebin. “That brings heavier traffic, as well as heavy goods vehicles, onto a road that is not designed to cope with them.”
His predecessor, Gerard Le Gall, had commissioned road repairs and traffic-calming measures to try to make the public space safer. He also threatened to ban heavy goods traffic. But none of these solutions is sufficient against the constant flow of vehicles that keep being redirected towards Tremuson. In Mr Orgebin’s opinion, applications are “a bit too fast at pushing drivers towards secondary roads. I’ve seen it suggested when the main road was not even really congested.” The mayor's office has tried to reach out to some GPS applications, he says, without success.
Waze, an archetype of a wider problem
Because Waze puts the emphasis on its mission to avoid slowdowns, the people AlgorithmWatch interviewed criticise the brand more specifically than other players. “One of its problems is that it uses its own classification, which differs from the French classification of its highways, roads, and streets,” Mr Combe says. For instance, urban streets leading to schools or hospitals “shouldn’t attract heavy traffic. But Waze is blind to the difference between these and traditional urban streets.”
But Waze (used by 1.3 million French people every day in 2022), Google Maps (2.2 million daily French users), or Apple Plans (1.2 million French users) all provide a similar service. They do “nothing illegal”, says Antoine Cormont. “They just expand a practice that previously existed of using shortcuts and so-called ‘clever itineraries’.”
Not much is known about the exact functioning of the algorithms that calculate the routes on each of these applications, apart from what Google and Apple share themselves. Thanks to its open editing community, however, a few things can be said about the way Waze works - even Waze’s editors themselves speculate about the technology on their forums.
Anyone can participate in maintaining and upgrading the quality of Waze’s map, with editing rights increasing with usage. The most important feature to which these user-editors contribute is road classification: a route can be described as a freeway, a major highway, a minor highway, a primary street, or a street. It’s on this basis, and using all the other data that editors and users pour in (roadworks or change of traffic direction, for example), that Waze runs its calculation.
“Waze uses an A* path search algorithm which allows calculating the shortest route from an A to a B point,” says Matthieu Casanova, who started editing the applications’ map back in 2010. He admits he does not know much more, but experience taught him some details, like the fact that compulsory speed is not included among the data used to calculate the fastest route.
“This has created friction between the user-editors and the Waze teams, because in some cases, while driving at 130 km/h on a highway, you would still see your estimated arrival time get delayed. If that happens, it means that the average speed the algorithm uses to calculate your travel time is above the legal limitation,” he adds.
A feedback loop into offline life
Still, in many places in the world, infrastructure managers and mayors irritated by the traffic displacement towards secondary roads and streets have taken action, creating a feedback loop of sorts from the digital tools. In the United States, for instance, a few cities have entirely banned traffic from non-residents. In France, an often-used solution consists in reducing the speed from 50 km/h to 30 km/h within portions of cities.
“They can also create chicanes or issue bylaws to limit tonnage,” says Christophe Lautier, an editor since 2013 and the French community partnership manager (a voluntary position) for Waze. “If the speed statistics get lower and lower, the algorithm will eventually pick up that it's not so advantageous to use these roads.” But that might be too self-serving an attitude: according to the sociologist Mr Courmont, “some municipalities are worried that, if they change their traffic plan too drastically, it could impact the neighbouring town”.
In charge of “a day-to-day dialogue with town halls, the police, event teams, and others”, Christophe Lautier says there is always room for discussion. He brings up the example of Rungis. Even though it holds the Rungis International Market, the second-largest wholesale food market in the world, “it is a relatively small town [around 5,600 inhabitants], which does not give them much negotiation power with the authorities”.
Recently, the mayor called him, explaining they had a crossroads filled with traffic jams every morning because of the market. “We looked at Waze’s map together, and indeed, we discovered a miscategorization of streets that did not take into account the new road bypassing Rungis.”
According to the editor, this type of collaborative work helps everyone: cities and motorway operators can declare events, roadworks, or other elements likely to disrupt traffic, “and, in return, they get data regarding average speed, potholes, accidents, signs that don't work, etc”. Six months after the beginning of this cooperation, according to Rungis’ municipal councilor Martin Jardillier, “some residents and retailers see a positive change, others feel that it has changed nothing.”.
A national database to help cities regain control
In fact, the navigation application created a program to help scale this type of collaboration: Waze for Cities. Mr Lautier finds it positive in every way, as it also helps Waze editors keep their map as accurate as possible.
According to Mr Courmont’s work, however, such cooperation is not as impactful as the disturbance that GPS navigation systems create in the daily life of relief roads’ residents. In Rungis, Mr Jardillier sees how the program can prove useful, but he believes “it must work better for bigger cities”. A small town like his own, he explains, rarely has “the capacity to dedicate the part-time or full-time work that would allow using the platform to its full potential”.
To help them, another solution has been devised in France, a legal one. Implemented in the wake of the pandemic, the decree states that competent authorities can reclassify sections of secondary roads that should not be used as relief roads for primary routes. They should be able to do so directly in the “BD Topo”, a national database managed by the French Nation Institute for Geographic and Forestry Information (IGN) in conjunction with the Ministry for Transport.
With this tool, municipalities will be able to identify which sections can be used as relief routes in the event of traffic jams, and which cannot – and anyone, navigation applications included, will be able to query the database to obtain this information. “It will serve as a one-stop-shop both for local authorities, who otherwise have to try and deal with each and every GPS application, and for tech firms, who otherwise would be supposed to check with each and every player to avoid discontent,” says Matthieu Le Masson, who manages the project for IGN. Time will tell to what extent this will help.
Apple, Google, and Waze have been contacted for this piece. They did not provide comment.
Mathilde Saliou
Former Fellow Algorithmic Accountability Reporting
Mathilde is a French journalist specialized in digital issues. A graduate of Sciences Po Paris, she has worked for The Guardian, RFI, 20 Minutes, Les Inrocks, Next INpact, and others. In 2023, she published the book Technoféminisme, comment le numérique aggrave les inégalités ("Technofeminism, how digital technology is exacerbating inequalities") with the French publishing house Grasset.